NeoNet app and dataset
Mapping radiocarbon dates from Late Mesolithic/Early Neolithic transition in the North Central-Western Mediterranean watershed
Thomas Huet, Niccolo Mazzucco, Miriam Cubas, J. Francisco Gibaja, and F. Xavier Oms
Figure 1: Preview of the current NeoNet C14 dates dataset, in red, covering the Late Mesolithic/Early Neolithic transition of the North Central-Western Mediterranean watershed
Presentation
The NeoNet app aims to contribute to the study the pioneering front of the farming live-style (i.e., Neolithic) by focusing our study on the mobile border between the Last Hunter-Gathers economy and the Early Neolithic (ca. 7,000 BC to 3,500 BC) in the Central and Western Mediterranean. The interactive web app facilitates the selection of absolute dates (c14 dates) by providing selection tools for:
- spatial
- geographical region of interest (ROI)
- custom selection shape inside the ROI
- chronology
- date time span between a tpq and a taq in cal BC
- main periods (Middle Mesolithic, Late Mesolithic, Early Neolithic, …)
- date accuracy
- some type of material life duration (short like, long life or others)
- threshold of the maximum accepted standard deviation (SD)
This static website is a tutorial for the RShiny NeoNet app. Both have been generated with R, from a perspective of FAIR Science and collaborative working. The app has been developed in the frame of the NeoNet work group. We will see how it works, what is the format dataset and what are our objectives
NeoNet app
The app is a RShiny hosted on the Università di Pisa server. The app is divided into five (5) panels:
- map panel: spatial filtering with selection menus on dates
- calib panel: calibration of the selected dates
- data panel: the whole dataset
- biblio panel: bibliographical references
- infos panel: credits and link to the webpage handbook of the app
1. map panel
The panel map is a geographical window provided by the Leaflet package. This panel is used for selection of radiocarbon dates by location, by chronology, by quality of dates. Once selected, dates can be calibrated
Figure 2: The different menus of the map panel
The current functions are:
Fig. 2, red box, top-left button group C14 on map: allows to cluster dates by spatial proximities (Marker Clusters)
Fig. 2, pink box, top-right layer button: allows to change the basemap. By default, the basemap is OSM, an OpenStreetMap general basemap, but it can be switch to Topo, an ESRI topographical basemap
Fig. 2, grey box, bottom inline text: reactive count of selected dates and select sites
Fig. 2, green box, bottom table: reactive datatable (DT package) listing all the dates within the map extent (ROI) and the optional selection menus (tpq/taq, material life duration, maximum SD, periods, selection shapes)
select by location
By default only the data within the window extent (ROI) will be selected. But selection shapes can be drawn inside this ROI to have a spatial intersection:
Fig. 2, black box, top-left draw toolbar: selection shapes, polygons and rectanges, can freeze the date selection inside a given ROI. They can be removed with the trash button. All the dates inside the ROI and selected with the others filters will be visible on the map, but only those inside the selections shapes will be calibrated
selection inside a shape, here a single polygon. Before shape selection: 190 sites and 895 dates. After shape selection: 13 sites and 68 dates
retrieve coordinates from the map
As said, the default basemap of the app is OSM. It offers a well documented basemap where archaeological sites are sometimes already located, like the Ligurian site of Grotta della Pollera. Clicking on the map show the lat/long coordinates of the current point (under the tpq/tap slider). These coordinates can then be copied and used to modify the NeoNet dataset
get coordinates by clicking on the map
select by chronology
select by dates quality
- Fig. 2, purple box, bottom-right checkboxes and slider: a group of menus for selection on the material life duration and max accepted SD:
- relatively to the duration of their material (short to long-life material)
- below a maximum accepted threshold for the standard deviations (SD) for the dates
calibrate one or various dates
The dates displayed in the table of the map panel will be calibrate when one of them has been clicked.
click on a date to calibrate a selected group of dates
2. calib panel
The panel calib is used for analysis. Calibration of selected dates are done on-the-fly with the R package rcarbon. If the dates are numerous (e.g., > 100) the computing time could take times, be patient.
calibration of selected radiocarbon dates
This date which have been clicked on the map panel will be shown bolded on the output figure
c14 group by filter
The only selection which can be done is on the top-center radio button (Fig. 3, red box). The c14 group by filter allows to plot dates and to sum their probability densities depending on different levels of grouping:
LabCode: each radiocarbon date is plot separately (by default)
SiteName and PhaseCode: dates from the same site, having the same archaeological unit (layer, structure, etc.), are summed. See the PhaseCode field.
SiteName and Period: dates from the same site, having the same period are summed
Period: dates having the same period are summed
All C14: all dates are summed
Figure 3: The different menus of the calib panel
3. data panel
The complete dataset from the GitHub c14_dataset.tsv file (a dataframe with tab-separated values). Today, the dataset counts 2506 dates:
| SiteName | Country | Period | PhaseCode | LabCode | C14Age | C14SD | Material | MaterialSpecies | tpq | taq | Longitude | Latitude | bib | bib_url | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1804 | Peiro Signado | France | EN | Structure 1 | Ly-8400 | 6840 | 55 | wood charcoal | n/d | -5838 | -5631 | 3.328910 | 43.30830 | Binder et al. 2018 | https://doi.org/10.4312/dp.44.4 |
| 1061 | Cueva de Fuente Hoz | Spain | MN | Layer Ia et b | I-11588 | 5240 | 110 | human bone | Homo sapiens | -4331 | -3801 | -2.893699 | 42.80537 | Sarasketa-Gartzia et al. 2018 | https://doi.org/10.1007/s12520-018-0678-7 |
| 805 | Les Plots | France | MN | n/a | Gif-9336 | 5170 | 50 | wood charcoal | n/d | -4220 | -3801 | 2.404269 | 43.22056 | Vaquer 1998 | Vaquer98 |
| 1933 | Costamar | Spain | EN | n/a | n/a-4 | 5996 | 38 | animal bone | Bos taurus | -4994 | -4792 | 0.158587 | 40.13683 | Ibanez et al. 2017 | https://doi.org/10.1007/978-3-319-52939-4_13 |
At first, these data were recorded in an Excel spreadsheet (c14 spreadsheet) in order to facilitate their editing (filter, sorting, fill). Then a R function reads this spreadsheet, calculates the tpq and taq (in cal BC) of each uncal BP (with Bchron), and converts it into a .tsv file. As data came from various publications, a homogenization the different values (material, cultures, bibliographical references, etc.) has been done. The dataset mandatory fields are:
- SiteName: the site name
- coordinates (two fields)
- Longitude: in decimal degrees (ex:
1.0451) - Latitude: in decimal degrees (ex:
43.9213) - Period: a value as following: LM (Late Mesolithic), UM (Undefined Mesolithic), LMEN (Late Mesolithic/Early Neolithic), EN (Early Neolithic), MN (Middle Neolithic), LN (Late Neolithic), UN (Undefined Neolithic)
- PhaseCode: a code for the dating stratigaphical unit and/or structure
- C14Age: a numerical radiocarbon dating in BP
- C14SD: the standard deviation (SD) of the radiocarbon dating
- LabCode: the unique identifier of the radiocarbon dating
- Material: a value from the material.life spreadsheet
- MaterialSpecies: a specification of the field Material
- tpq: the terminus post quem of the radiocarbon dating in cal BC
- taq: the terminus ante quem of the radiocarbon dating in cal BC
- bibliographical references (two fields)
The recommended fields are:
- Culture: a specification of the field Period
The others fields (if there’s any) only concern the EUROEVOL_R app
mandatory fields
Here we explain more precisely some of the mandatory fields
Longitude and Latitude
In in decimal degrees and a precision of four (4) decimal digits (ex: 1.0453, 43.9211). Since the modeling is supra-regional, the app does not need to record dates with high accuracy geographical coordinates. At the minimum, this accuracy can be a location inside the departmental/county boundaries (how to retrieve better coordinates from the map)
PhaseCode
The PhaseCode field provide a more precise archaeological context than the site name. It is useful for layer/structure C14 grouping. Most of the time, it correspond to an archaeological layer or structure
| PhaseCode |
|---|
| C5 |
| C7-8 |
| foyer 7 |
| niv. II |
| … |
Values of this field need to be homogeneized (for example: C.5 or layer 5 -> C5) – at first for the same sites, in order to group them – then for the whole dataset. The n/a value (i.e., not available) is reserved to dates without intra-site contextual information
LabCode
LabCode (i.e., laboratory code) should be unique. Their conventional syntax is ‘AbrevLab-number’, respecting the case letters (upper case and lower case). For example:
| LabCode |
|---|
| Beta-103487 |
| CSIC-1133 |
| ETH-15984 |
| Gif-1855 |
| GrN-6706 |
| KIA-21356 |
| LTL-13440A |
| Ly-11338 |
| MC-2145 |
| OxA-9217 |
| Poz-18393 |
| … |
See also the list of laboratories. Exceptionally, if a date has no LabCode – e.g., the ‘Sep-H3 mix’ from Fontbregoua, 6082 +/- 35 BP – the convention is to use the PhaseCode (e.g., ‘Sep-H3 mix’) with an underscore as a prefix (e.g., ’_Sep-H3 mix’) to get an unique key.
bib and bib_url
Every radiocarbon date should be referenced with a short plain text bibliographical reference (bib field) and a DOI, or a BibTex key (bib_url field). We favor the earliest mention of the radiocarbon date.
bib
The plain text that will be plot for each radiocarbon date under the bibliographical reference section. Basically the name of the author(s) and the publication year, for example Guilaine et al. 1993, Binder 2018 or Manen et Sabatier 2013. The values of this field can be the same for two different publications (e.g. Delibrias et al. 1982 refers to two different publications the same year)
bib_url
Either a DOI (starting with “10”) or a unique BibTeX key. We favor the DOI as a unique bibliographical reference. The values of this field should be unique for a single publication (e.g. the BibTeX keys Delibrias82 and Delibrias82a). For example:
| tpq | taq | select | RedNeo | bib | bib_url |
|---|---|---|---|---|---|
| -6086 | -5923 | VRAI | 1 | Binder et al. 2018 | https://doi.org/10.4312/dp.44.4 |
| … | … | … | … | … | … |
When the DOI is lacking, the bibliographical reference should be include into the BibTex document id00140_doc_reference.bib with the name of the first author and the two last digits of the year:
@book{Guilaine93,
title={Dourgne: derniers chasseurs-collecteurs et premiers {\'e}leveurs de la Haute-Vall{\'e}e de l'Aude},
author={Guilaine, Jean and Barbaza, Michel},
year={1993},
publisher={Centre d'anthropologie des soci{\'e}t{\'e}s rurales; Arch{\'e}ologie en Terre d'Aude}
}The key of this reference is added to the bib_url field. For example, the key value Guilaine93 from the c14 spreadsheet will match this complete reference
| tpq | taq | select | RedNeo | bib | bib_url |
|---|---|---|---|---|---|
| -3330 | -2492 | FAUX | 1 | Guilaine et al. 1993 | Guilaine93 |
| … | … | … | … | … | … |
correspondance tables
The NeoNet app makes joins to two tables in order to retrieve information and to provide a handy user interface
material.life
Material life duration are read from the GitHub c14_material_life.tsv file. The two fields show the material type (column 1) and the material life duration (column 2), for example:
| material.type | life.duration |
|---|---|
| Drusch - trilladura | long.life |
| fauna | short.life |
| Agla | long.life |
| Angiosperma | long.life |
| animal bone | short.life |
| animal hair | short.life |
This thesaurus is used both by the NeoNet app and the EUROEVOL_R app. Specifically, the NeoNet uses these values:
| material.type | description |
|---|---|
| CE | Cerealia |
| F | Fauna |
| H | Human |
| OR | Organic |
| SE | Seed |
| SH | Shell |
| WC | Wood charchoal |
In the NeoNet app, this dataset is rendered with the (DT package) allowing sorting and filtering tools
4. biblio panel
Bibliographical references are recorded in id00140_doc_reference.bib file. If only exist a BibTeX key, and no DOI, this file results of the join between the bib_url field of the C14 spreadsheet and the id00140_doc_reference.bib.
## Warning in doTryCatch(return(expr), name, parentenv, handler):
## id00140_doc_reference.bib:33:6
## syntax error, unexpected TOKEN_VALUE, expecting TOKEN_EQUALS
## Dropping the entry `Forenbaher15` (starting at line 31)## @Article{Binder18,
## title = {Modelling the earliest north-western dispersal of Mediterranean Impressed Wares: new dates and Bayesian chronological model},
## author = {Didier Binder and Philippe Lanos and Lucia Angeli and Louise Gomart and Jean Guilaine and Claire Manen and Roberto Maggi and Italo M Muntoni and Chiara Panelli and Giovanna Radi and others},
## journal = {Documenta praehistorica},
## volume = {44},
## pages = {54-77},
## year = {2018},
## publisher = {University of Ljubljana Department of Archaeology},
## }
##
## @InProceedings{Briois09,
## title = {L'abri de Buholoup: de l'{\'E}pipal{\'e}olithique au N{\'e}olithique ancien dans le piedmont central des Pyr{\'e}n{\'e}es},
## author = {François Briois and Jean Vaquer},
## booktitle = {De M{\'e}diterran{\'e}e et d'ailleurs...: m{\'e}langes offerts {\`a} Jean Guilaine},
## pages = {141-150},
## year = {2009},
## }This BibTex file is rendered in HTML with an APA citation format (field long.ref) with the read.bib() and the markdown() functions (among others)
## Binder D, Lanos P, Angeli L, Gomart L, Guilaine J, Manen C, Maggi R,
## Muntoni IM, Panelli C, Radi G, others (2018). "Modelling the earliest
## north-western dispersal of Mediterranean Impressed Wares: new dates and
## Bayesian chronological model." _Documenta praehistorica_, *44*, 54-77.
##
## Briois F, Vaquer J (2009). "L'abri de Buholoup: de l'Épipaléolithique
## au Néolithique ancien dans le piedmont central des Pyrénées." In _De
## Méditerranée et d'ailleurs...: mélanges offerts à Jean Guilaine_,
## 141-150.
##
## Oms FX (2008). "Caracterizació técnica, tipológica i cronológica de les
## ceràmiques del Neol\'itic antic de la Cova Colomera (Prepirineu de
## Lleida)." _Archivo de Prehistoria Levantina_, 51-80.
##
## Komšo D (2008). "Mezolitik u Hrvatskoj." _Opusc. Archaeol. Rad. Arheol.
## zavoda_, *30*, 55-92.
Collaborative work
NeoNet workgroup, and NeoNet app, aim to facilitate contributions in a perspective of FAIR Science. In practical terms, we have planned to:
- host the app and dataset on an institutional web server
- publish the dataset in an Open Data repository to get a DOI
- submit the dataset in a data paper (JOAD)
- create a connector to the dataset with the c14bazAAR getter function
- open the app to new contributions
- publish the app source code in an Open digital humanities paper (ex: JOSS)
Developments
Currently, the NeoNet database and app received the contributions of the NeoNet workgroup collaborators:
- Thomas Huet thomashuet7@gmail.com
- Niccolò Mazzucco nicco.mazzucco@gmail.com
- Miriam Cubas mcubas.morera@gmail.com
- Juan Francisco Gibaja jfgibaja@imf.csic.es
- F. Xavier Oms oms@ub.edu
The development version of the app is on GitHub: zoometh/neonet. You can check the contribution rules and the relevant license.
Regarding the app, we expect to integrate the Bayesian modeling of PhaseCode during the calibration process (calib panel) with RChronoModel functions. Regarding the dataset, we expect to record the radiocarbon dates of the Middle and Southern European Atlantic watershed (ie. western Iberian peninsula and Western France)
| SiteName | Country | Period | PhaseCode | LabCode | C14Age | C14SD | Material |
|---|---|---|---|---|---|---|---|
| Grotta delle Arene Candide | Italy | EN | Phase AC1, US360 | LTL-15944A | 6864 | 45 | animal bone |
| Grotta S. Croce | Italy | EN | US 78 | OxA-7596 | 6555 | 50 | plant seed |
| Cova de Can Sadurni | Spain | MN | US 11b | Beta-210653 | 5790 | 40 | animal bone |
| Cova de Can Sadurni | Spain | EN | US 19 | Beta-398967 | 6400 | 30 | human bone |
| Villandro/Villanders | Italy | EN | US 30 tg 9 | ETH-30059 | 6105 | 60 | wood charcoal |